EFI AnalyticsData driven Tuning

MegaMeet 2014

Phil Tobin aka LT401VetteAndy Whittle aka whittlebeastBrian Holzbach aka Brian EFISept20, 2014

MegaSquirt

Overview

Getting started – Basic tuning tips

Review key configuration setting

General Tuning rules Explanation of the different areas of the tune Common settings that can skew Ve Table Values What to look at before you begin tuning Check timing table values & tune Spark Table Tuning Ve Table Analyzing Data Logs Table Smoothing Cranking, ASE and Warmup Enrichment

EFI Analytics , Inc

Getting started - Tuning Tips

EFI Analytics , Inc

Tuning is a process you must be patient

Always data log while making tune changes

Use data logs to find what your engine wants

95% of the time the answers are in the data !

Concentrate on one area of your tune at a time

Only make one or two changes at a time (very easy to loose track)

Save settings often in case you need to revert back

Avoid lean AFR readings, especially under load

Conservative timing advance (detonation kills engines & drains wallets)

Before you begin tuning

EFI Analytics , Inc

Look over your settings

Research what you do not know

Verify Sensor calibrate

O2 Correction

Before you begin tuning

EFI Analytics , Inc

Settings that can dramatically alter your VE Table

Injector dead time

Incorporate AFR target

Multiply MAP

General Rules

Acceleration Increase fueling to achieve AFR target

Increased timing advance full timing around 3500

Best Throttle response

EFI Analytics , Inc

Accel enrichment options TPS MAP X-Tau EAE

AFR Table

EFI Analytics , Inc

I

Best values will vary some with engine Best targets will vary with Wideband accuracy AFR Table Generator will provide a good starting point Use dyno or ET’s to find best WOT AFR Use fuel economy and best running for Cruise AFR Use best idle for Idle AFR

Idle – generally stoich or slightly rich

Cruise – Stoich to slightly lean

WOT / Boost – rich of stoich.• NA – 12:1 – 13:1• Boost richen more as boost

increases

Idle, Cruise & WOT

Fairly rich (12.5:1 to 13.5:1 ) boosted applications will (11.5:1 to 12.1:1)

EFI Analytics , Inc

Stoichiometric to 15.5 AFR

High timing advance to full time

Typically Stoichiometric to 15.8:1 but can vary with application

Roughly 5° to 15° BTDC

Generating Base line advance table

Start with factory advance table if possible

(NA) reduce whole advance table by 2°- 3°

(Boosted) reduce by 2°- 3° / above 100kPa

reduce an additional 1°- 2°

Generating advance Table based on engine

parameters

older engines (1960s to 1990) with two valves -

max advance = 36

Newer two-valve engines – max advance = 30°

Three or four valve engines – max advance = 26°

Adjust for bore size

Under 3.5” (89mm) – subtract 3°

Between 3.5” & 4.0” (101.6mm) – no adjustment

Over 4.001” (+101.6mm) – add 3°

Adjust for the fuel

Regular – subtract 2°

Mid-grade – subtract 1°

Premium – no adjustment

Additional Modifiers

Aftermarket combination good squish +

optimized quench - subtract 2°

Idle to 3000RPM @ 100kPa increase

advance linearly to max advance

Every 10kPa over 100kPa reduce advance

- .5° to 1°

Max advance at full boost roughly half of

your total adv. at 100kPa

These are rules of thumb to help create a conservative spark advance table to get the engine runnable. The base line spark table you made from this information will need fine tuned to make optimal power.

Ve Table Tuning

EFI Analytics , Inc

Ve Table Tuning

EFI Analytics , Inc

I

MegaSquirt firmware uses 4 cell interpolation Good Ve Table Values guideline 30 – 130

Values typically increase as you climb through the table

If VE Values out of range adjust required fuel

recommendedVE = currentVE * (egoCorrection/100) * (actualAFR / targetAFR)

Typical Values: Idle 30 – 60

Cruise 50 – 80

WOT 80 -130

Boost 85 – 140

Decel 30 - 50

Ve Table Tuning

EFI Analytics , Inc

I

VE Analyze

Log based or Live in TS

Must be able to identify when something isn’t right

Manually adjusting areas Ve Analyze live did not reach

Some sections of the table will not get enough or any data to be corrected.

These areas of the table are not used while running the engine but you would likely want to manually correct

Cranking

EFI Analytics , Inc

Increase fueling for cold conditions

Low Ign advance (minimize kick back)

Cranking Pulse Width

Adjust for best start at temp

ASE – After Start Enrichment Additional fuel for a short

period after start Lean immediately after starting Data log of start up best to

identify after start lean Hot engines need less

Warm - Up

EFI Analytics , Inc

WUE - Warm Up Enrichment Increased Fuel based on given conditions

Increased timing advance

Warm up enrichment

ASE – After Start Enrichment Additional fuel for a short period

after start Lean immediately after starting Data log of start up best to identify

after start lean Hot engines need less

WUE with Scatter Plots

EFI Analytics , Inc

Correcting Warmup Enrichment with Scatter PlotsFull view for less iterations

CLT vs Gwarm with SmoothAFR for the Z Axis color• Regenerate the Warmup Curve as it was set in TunerStudio• Notice the Z Axis color that shows how far from target AFR it was at each

temperature.

SmoothAFRError? What is that?Custom Field explained on next slide

Solving WUE

EFI Analytics , Inc

Not only can you view what the warmup curve looked like, but you can generate the “Right” answer for it.

Note the Z Axis color illustrates the original Warm up

enrichment value.

The X & Y Axis show the desirable value.

This car was running needlessly rich.

How?? Where did the new curve come from?Desired Warm up for a single point would be:

Desired Warmup = [Gwarm] * ([Gego]/100) * ([AFR]/[targetAFR])Add Smoothing:

Desired Warmup = [Gwarm] * (SmoothBasic([Gego], 20)/100) * (SmoothBasic([AFR], 20)/[targetAFR])

Analyzing the data to know what changes need made to your tune.

EFI Analytics , Inc

Reducing condition changes Methods of looking at data Ways to spot issues

MegaLogViewer Scatter Plots

EFI Analytics , Inc

Abstract Art or valuable tuning tool?

Common Quote: Sure it looks pretty, but what the hell do I

do with it? What is it telling me?

MegaLogViewer Scatter Plots

EFI Analytics , Inc

Overview: Scatter Plot graph are commonly used for statistical and data analysis to identity correlations

MegaLogViewer Scatter Plots were designed for Tuning• Handles high data volumes• 3 Axis view • Custom Calculated Fields to combine and manipulate• Data Filtering!!!

Common Belief – For tuning high data rates are very important.FALSE – for most tuning slow data rates will work fine as long as each record is well correlated and you have sufficient time to collect data. More data is better with Scatter Plots, not necessarily faster data.

Normal log viewing work great for viewing specific details, but you can only see so much data at once. Using scatter plots you can view thousands of records or hours of data in 1 view to quickly get a big picture view.

MegaLogViewer Scatter Plots

EFI Analytics , Inc

When a scatter plot over a traditional graph?

- View many fields at once

- Step through record by record to see specifically what was happening at any point in time.

- Can see that at this specific point in time the RPM=5612, MAP=93 kPa, AFR is 12.55:1 and

Ego Cor is 100

- Difficult to realize the impact of conditions not happening at other points in time are having on the

viewed fields

- 3 Fields at a time

- Full picture view as if it is your Speed density VE Table. X Axis:RPM, Y Axis: MAP, Z Axis:AFR

- Notice the Yellow spots, they quickly reveal areas of the VE Table where it is running over 16:1. This is a

trend over a large amount of data, not a specific point.

- Can peel back the layers with filters to see what conditions may be effecting the view. AE, Warmup.

Quick View of Tune

EFI Analytics , Inc

So what? - MAPxRPM is just that: [MAP] * [RPM]

- Compare to DutyCycle or MAF on the Y AxisThis is really fast way to spot a rough tune or better, where the problems are, one quick view.

Calculated Field MAPxRPM vs DutyCycle – gives tight correlation on a well tuned engine.

Quick View of Tune

EFI Analytics , Inc

Let’s look at this log in MegaLogViewer.- Bring up the Scatter Graph MAPxRPM vs DutyCycle1

- The Dead O2 Filter removes some records in the hump, but an insignificant number.- Click on the hump

- Switch to normal log viewing- Observe what is happening on many of those records.

The scatter plot showed there was a problem and where to looks. The normal log shows specifics.

Not showing a clear correlation.

Why?

What looks wrong?

What is causing this hump?

Scatter Plot Filters

EFI Analytics , Inc

Filtering:

Back to our CLT vs Gwarm with SmoothAFR for the Z Axis color

Filters can quickly remove anomalous data, or help to find the cause of that data.

Notice the blue dots up the left axis of top picture.Activate the dead O2 Filter to remove in picture 2.

Now you know you would want to filter that data from any formulas that it would skew the results.

Dead O2 Filter expression with Innovate 10:1-20:1:[Field.AFR] < 10.3

Numerous built in Filters where you can see the expressions used:

TP Transient: (abs([Field.TP] - [TP-4]) > 3 ) ASE Active: [Engine] | 4 == [Engine]

MAP Transients: (abs([Field.MAP] - [MAP-4]) > ([Field.MAP] * 0.15) )

These transient filters are built in and available for use after install.

EFI Analytics , Inc

Questions?

Hiring

Interested in working with the MegaSquirt team?EFI Analytics is hiring

- Tech Support- Java Developer

Please catch up with me here today if interested or email me

EFI Analytics , Inc